andrewavit.bsky.social
Education Policy PhD student at UVA
Evaluating education policies, with a particular focus on teachers, in the US and LMICs
https://www.andrew-avitabile.com
86 posts
306 followers
315 following
Regular Contributor
Active Commenter
comment in response to
post
Out of curiosity, how do you think policy evaluation work falls into all this? I feel like we have a good base of evidence on things like performance pay, but probably still worth evaluating as we implement things. Maybe it’s just a matter of not overselling contributions?
comment in response to
post
Leave my grandpa out of your housing policy 👴🏼
comment in response to
post
All their dairy products are really good, but imo their parmesan is the best I’ve had from a grocery store 🧀
comment in response to
post
It should at least get you pre-check
comment in response to
post
[12/12]
If you're interested in teacher ed, K-12 ed policy, or exploring LLMs in research—this paper is for you!
📄 Read here: edworkingpapers.com/ai25-1203
We'd love your feedback!
comment in response to
post
[11/12]
We also show how LLMs can help analyze and code large text datasets. Our method reliably extracts feedback elements and teaching skills—replicable, fast, and validated against human coders.
If you're interested, I’ve got a toy example available on my github: github.com/avitabilea/C...
comment in response to
post
[10/12]
Key Takeaways:
1️⃣Feedback during the clinical teaching experience quality is uneven and often low-quality
2️⃣ PST skill gaps—esp. classroom management—predict lower scores and weaker job outcomes
3️⃣ Low-cost fixes like better prompts or supervisor training could help a lot
comment in response to
post
[9/12]
We also find that PST skill gaps are related to more distal labor market outcomes:
1️⃣ PSTs flagged for lesson planning score worse on certification exams
2️⃣ PSTs flagged for classroom management or student engagement are less likely to return to their placement school as teachers.
comment in response to
post
[8/12]
These areas for improvement are related to PST clinical performance:
📉 Classroom management feedback → lower observation scores
📈 Differentiation feedback → higher scores
comment in response to
post
[7/12]
When feedback includes an area for improvement (which it often doesn’t), supervisors and PSTs focus on different things:
🔶 Supervisors emphasize classroom management
🔷 PSTs focus on lesson planning
comment in response to
post
[6/12]
Feedback quality also varies considerably by…
1️⃣ Observation order
📉 Certification area
🚨 Placement school disciplinary climate
🧑🏫 Supervisor leniency
📊 Table 1 has the details, but I’ll spare you the regression output here.
comment in response to
post
[5/12]
Feedback quality isn’t random.
➡️ Some supervisors never give areas for improvement, while others always do.
📊 A big chunk of the variance in feedback quality comes down to who your supervisor is.
comment in response to
post
[4/12]
What do we find?
❌ Fewer than half of supervisor evaluations include areas for improvement or next steps.
✅ In contrast, PST reflections usually include all four quality indicators.
comment in response to
post
[3/12]
Do do this, we process 11,000+ supervisor evaluations & PST reflections using LLMs (i.e., ChatGPT) to flag:
✅ PST Strengths
✅ Specific examples
✅ Areas for improvement
✅ Actionable next steps
And when areas for improvements are flagged—which teaching skills are discussed?
comment in response to
post
[2/12]
Why does this matter?
Clinical teaching is critical for new teacher development. But while we focus a lot on mentor teachers, the role of program supervisors—who give feedback during field observations—is understudied. We help fill this gap.
comment in response to
post
This is how we’ll soon all be coding 🤯
comment in response to
post
Big congrats, Wes! 🎉
comment in response to
post
Not one, but folks at MDRC have compiled a dataset from over 25 RCTs with over 65000 individuals. More info here:
www.mdrc.org/work/project...
comment in response to
post
Congrats and welcome to UVA!!!
comment in response to
post
This has changed so frequently with the Temer-Bolsonaro-Lula presidencies… it’s always confusing. Apparently you can get a visa electronically these days:
www.gov.br/mre/pt-br/co...
comment in response to
post
I think Todd & Wolfpin (2003) formalize this, but I haven’t heard of a solution. Curious to hear what you and others come up with!
onlinelibrary.wiley.com/doi/abs/10.1...
comment in response to
post
Looks like a graduate student office
comment in response to
post
This work is super appreciated!
comment in response to
post
Congrats!!! 🍾
comment in response to
post
@iancallen.bsky.social
comment in response to
post
“Extreme fear” couldn’t be funnier though
comment in response to
post
I’ll ask around and see if folks have recommendations!
I also know @fenghu.bsky.social and @alexeble.bsky.social have a few education studies in China. Perhaps they can steer you in the right direction :)
comment in response to
post
What language are you translating to and for which country? There’s a network of ex-JPAL and IPA folks who might have more country-specific info. Would be happy to reach out on your behalf
comment in response to
post
Avoid comparing one call for a recession to an already call for recession
comment in response to
post
Thank you! I just told @iancallen.bsky.social you’re my hero.
I’m trying to look at the effect of school-level events in teachers so any data from the past 20 years are really helpful!
comment in response to
post
Feels like there's a specific story behind this one 👀🍵